Population: Has India Really Overtaken China?

China's lead grew for a quarter of a century and peaked in 1975, 916.1 million people against India's 611.3 million. It shrank slowly at first: between 1980 and 2002 the gap closed by 106.5 million, about 4.8 million a year, a pace that would have run out somewhere in the 2040s.

It did not take that long. The remaining 190 million closed between 2002 and 2022 at roughly 9.5 million a year, nearly twice as much ground in two fewer years. That acceleration is why the crossover happened when it did, and it came from both ends at once: India's huge cohorts of the 1980s and 1990s reaching childbearing age while China's births fell away.

The limit rose to two children from 2016 and to three on 31 May 2021. Neither move worked. China recorded 7.92 million births in 2025, the fewest since records began in 1949 and about 17% below 2024's 9.54 million, against 11.31 million deaths. The population fell 3.39 million to roughly 1.405 billion, a fourth consecutive decline and the steepest on record outside the famine of 1959-61.

That is what a settled age structure does. China peaked at 1,426.4 million in 2021, and the decline now running does not wait on anyone's decision about having a child: too few women are of childbearing age, and too many people are old enough to die.

India's total fertility rate was 1.9 in the Sample Registration System report for 2024, below the 2.1 that holds a population level, and the national figure hides a spread from 2.9 in Bihar to 1.2 in Delhi, with only six states still above replacement. India still added 12.7 million people between 2025 and 2026, growing at 0.87% a year while China shrinks at 0.22%.

The two facts are not in conflict. Population momentum is the growth already locked into an age structure, the US Census Bureau's term for the change inherent in how a population is distributed by age. India has 1,013 million people aged 15 to 64 and only 112 million aged 65 and over. Even at 1.9 children each, that many adults near childbearing age produce more births than so small an elderly population produces deaths. The arithmetic flips only when those cohorts themselves grow old, which the UN medium variant puts in the early 2060s, at a peak near 1,701 million.

India's last completed census was in 2011. The 2021 round was prepared, then postponed for COVID-19, and never held. A Gazette of India notification of 16 June 2025, issued under the Census Act 1948, sets the reference moment for the next one at midnight on 1 March 2027, with 1 October 2026 for Ladakh and the snow-bound districts of Jammu and Kashmir, Himachal Pradesh and Uttarakhand.

The first phase, the houselisting and housing census, was scheduled to have concluded between 1 April and 30 September 2026. Population enumeration follows in February 2027.

That is what the 243,647-person margin rests on: not a count but a cohort-component projection, rolled eleven years past the last headcount and updated with sample registration. John Wilmoth, director of the UN Population Division, said when the crossover was announced that the precise timing is not known and will never be known.

The announcement makes the point by itself. UN DESA dated the crossover to April 2023; in the same month UNFPA put India ahead at mid-2023. Two arms of one organisation, two dates for one event, and Wilmoth called the mismatch unfortunate. The date that has circulated ever since is the date of a press release.

Yi Fuxian, a researcher at the University of Wisconsin-Madison whose appointment is in obstetrics and gynecology rather than demography, argues that China has overstated its population for decades. The claimed error moves between retellings, from more than 100 million in one interview to about 140 million where he puts the real 2020 figure near 1.27 billion against a census count of 1.412 billion.

His evidence is indirect: 17.7 million births recorded for 2000 against only 14.2 million lower-secondary graduates in 2015 despite compulsory schooling, and about 5.8 million BCG vaccine vials distributed a year in 2018-2020 where reported births should need nine to thirteen million.

Named demographers reject it. Feng Wang of UC Irvine says not a single person has discovered a discrepancy on that scale, and that three censuses since 2000 show no sign of one. Patrick Gerland, who heads the UN's population estimates section, says the UN does not take Chinese statistics at face value. No published paper has replicated or refuted the method, which is itself informative: the argument has run in newspapers, including the Wall Street Journal in October 2026, and not in the literature.

The documented failure runs the other way. China's own 2021 statistical yearbook put births for 2000 to 2010 at 172.5 million against the 160.9 million the 2010 census found in that age group, a shortfall of 11.6 million children, roughly 57% of them girls, hidden by parents until school registration while birth-planning penalties applied. Daniel Goodkind estimated in Demography that about 19% of children under five went unreported in 2000. The known bias is undercounting, and it has been corrected upwards more than once.

India passed China on working-age adults in 2024, 990.0 million to 984.0 million, two years after the totals crossed. China's working-age population had peaked at 1,000.0 million back in 2015, six years before its total population did.

The distance shows in the profiles. India's median age in 2026 is 29.2 years against China's 40.6, and India carries 11.1 people aged 65 and over for every hundred of working age where China carries 21.6. On the UN's medium scenario China's working-age population falls to 745.3 million by 2050, 255 million below its peak, with 52.2 elderly per hundred workers and 30.9% of the country over 65; India would have 1,134.1 million working-age people and 21.7 elderly per hundred. That divergence is what the crossover marks.

The dashed years are not measurements. Everything from 2024 is the UN's medium variant, one scenario among the several it publishes, and it is already drifting from China's own accounting: the UN has China at 1,416.1 million at mid-2025 while the National Bureau of Statistics reported about 1,405 million at the end of that year. Some of those eleven million is definitional, a mid-year estimate set against a year-end count, but not all of it.

Other models disagree more deeply. A Lancet analysis has India peaking near 1.6 billion in 2048 and falling to about 1.09 billion by 2100, where the UN has 1.70 billion in the early 2060s and 1.51 billion in 2100, roughly 400 million apart at the century's end. The crossover date is revisable too: it belongs to the 2024 revision, the UN revises about every two years, and the next one could move 2022 by a year in either direction.